“What Causes Muslims to Doubt Islam? A Quantitative Analysis”
February 13, 2018
Disclaimer: The views, opinions, findings, and conclusions expressed in these papers and articles are strictly those of the authors. Furthermore, Yaqeen does not endorse any of the personal views of the authors on any platform. Our team is diverse on all fronts, allowing for constant, enriching dialogue that helps us produce high-quality research.
Introduction
What motivates religious doubt among American Muslims? Answers to this vital question often rest on isolated anecdotes and intuition. At a time when the US population as a whole is becoming less and less religious, however, the need for a more systematic assessment of doubt in the American Muslim community is especially pressing. To address this demand, we sampled over 600 Muslims across America and recorded their opinions on a number of social, political, and religious issues.
Drawing on data from the inaugural Muslim American Attitudes Survey, this study offers the first quantitative insights on religious doubt among American Muslims and complements the qualitative findings reported in “Modern Pathways to Doubt in Islam.”[1] The following analysis is designed to both guide further research on this topic and provide a useable reference for imams and other counselors. First, it presents a descriptive breakdown of the extent to which various issues cause Muslims to doubt their faith. These summary statistics provide a means to compare the impressions of community leaders in “Modern Pathways” with American Muslims’ actual attitudes and experiences. Second, examining the determinants of religious doubt, this study highlights a number of positive and negative correlates, commenting on the (at times, surprising) associations the analysis reveals. The study concludes with an assessment of the broader project’s main contributions thus far and a look ahead to the research avenues that remain.
The Roads to and from Doubt
Methodology
Although this study breaks new empirical ground, the data collection expressly built on prior work when possible. Appendix A at the end of this report outlines the sampling technique in detail, however the sources of our questionnaire are worth highlighting at the outset. Specifically, the conceptualization and categorization of doubt aligns with this project’s initial study. As in “Modern Pathways,” here we are concerned with the type of religious doubt that potentially undermines one’s faith and could even lead one to abandon Islam altogether. To that end, the prompt that introduced the items comprising our doubt scale read as follows:
Sometimes, certain experiences or teachings lead people to deeply question their faith. To what extent have the following issues EVER caused you to seriously doubt your religious beliefs?
The doubt scale’s organizing framework similarly drew upon the findings from our previous study. In particular, this earlier work highlighted three core umbrella categories of doubt. The first, Moral and Social Concerns, reflects the potential anxieties that arise when an individual must reconcile their understanding of Islam with society’s (at times, far more malleable) ethical norms. It is worth underscoring that, contrary to anti-Muslim assertions that have gained prominence in the past decade (particularly since the start of the last presidential cycle), the overwhelming majority of Islamic doctrine and practice is wholly compatible with living as an American citizen. Yet, as with any comprehensive doctrine, there will be certain issues where mainstream interpretations of Muslim beliefs and practices are at odds with dominant cultural understandings in America. For American Muslims, these tensions and conflicts can become an impetus for doubt.
The second category, Philosophical and Scientific Concerns, comprises critiques aimed at the presumed “irrationality” of religion, generally, and Islam, in particular. Doubts that travel along this pathway tend to cluster around three nodes: 1) the Evolution vs. Creation debate, 2) the general perception that scientific knowledge conflicts with fundamental religious beliefs, and 3) the inability to “prove” certain tenets of faith or resolve seeming contradictions, such as why an all-Good God would allow evil in the world.
A final set of doubts, those stemming from Personal Trauma, rounds out the three general categories that emerged from in-depth interviews with Muslim institutional leaders across the country. This third source encompasses trauma brought on through 1) intimate events/interactions, whether acute (such as with the death of a loved one) or prolonged (such as with recurring physical or emotional abuse), and 2) communal interactions, wherein an individual feels discriminated against or otherwise unwelcome in dedicated Muslim spaces.
Utilizing this framework, we populated the doubt battery with items that other scholars have validated in previous studies. The collaborative and individual work of two sociologists, Bob Altemeyer and Bruce Hunsberger, was especially influential. In particular, a number of scales used in this study were adapted from the duo’s seminal, and entirely relevant, work, Amazing Conversions: Why Some Turn to Faith & Others Abandon Religion.[2] Extracting (with occasional necessary amendment) the most applicable items from Altemeyer and Hunsberger’s questionnaire and organizing the items according to our framework for Muslim doubt, we obtained the following battery of potential sources of doubt:
MORAL AND SOCIAL CONCERNS
PHILOSOPHICAL AND SCIENTIFIC CONCERNS
PERSONAL TRAUMA
For each item, the response choices were “Not at all,” “A little,” “A moderate amount,” “Quite a bit,” or “A great deal.”
Although the Muslim doubt scale largely aligns with our framework in “Modern Pathways,” there is one notable discrepancy: the items meant to gauge personal trauma do not ask directly about emotional or physical abuse. This was not an oversight, but rather a conscious decision. This determination was, first and foremost, taken with the concern of our respondents in mind, as such questions should not be broached absent a careful strategy to gauge the relevant attitudes and experiences while causing the least amount of harm. Such an approach requires a dedicated study on the topic, rather than one that measures a broad range of attitudes. Secondarily, including questions of this sort could be detrimental to the survey’s validity and reliability, as it could cause some respondents to simply drop out once they see these items, or otherwise alter the mindset of those respondents who choose to remain in the survey. For these reasons, we opted not to inquire about prolonged traumatic events in this survey.
The Sources of Doubt
Figure 1 presents the proportions for each of the items in the doubt scale, ordered from those issues that most caused respondents to deeply doubt their religious beliefs to those that did so the least. The first notable feature of this graph is that the top four issues all come from the Moral and Social Concerns family of doubts. Moreover, with the partial (though, debatable) exception of the top issue (“The way that [some] insist there is only one 'right' way to practice faith”), all four refer to particular behaviors rather than tenets of belief. The upshot is that the primary driver of doubt appears to be the actions of Muslims rather than the doctrines of Islam. This interpretation is further reinforced when turning to the bottom of the chart where each of the four items that least trigger doubt are directly related to doctrinal belief or the interpretation of personal life events in light of doctrinal belief.
To gain even more insight into the dynamics of Muslim doubt, we probed our respondents to consider the extent to which the issues that led them to question their religious teachings in the past are still a concern. Specifically, each respondent who noted that an item in the initial battery had caused them to seriously doubt their beliefs at least “A little” in the past, was asked about those same items again with the following prompt:
And of those issues that have troubled you in the past, how much do they CURRENTLY still cause you to doubt your religious beliefs?
Figure 2 presents the proportions for each of the items in the follow-up battery of doubt questions. Overall, our sample’s current level of doubt is significantly lower than in the past—a mean decrease of 23% on the overall doubt scale. Indeed, fully 17% of those who reported that at least one issue led them to doubt “A little” in the past now report that none of the items in our follow-up battery currently cause them to question their religious beliefs.
In terms of the sources of doubt, the top of the chart is once again dominated by moral and social concerns. The top two considerations driving current levels of doubt in our survey are “The bad things that people do in the name of religion” and “The problem of evil and unfair suffering in the world,” which indicates that negative headlines about terrorism and general hardship at home and abroad are taking their toll on the American Muslim psyche. The bottom of the chart remained largely unchanged from the pattern in Figure 1.
Responses to Doubt
Beyond cataloging the reported sources of doubt, we also wanted to examine how American Muslims respond when their faith is challenged. To answer this question, we gauged the likelihood of a respondent taking a particular action on a 4-point scale from “Not at all likely” to “Very likely” using the following prompt and survey battery:[3]
Thinking back on the instances when you DEEPLY questioned aspects of your religion or were troubled by certain beliefs or practices, how likely were you to do the following:
Figure 3 displays the proportions for the Responses to Doubt scale. The responses divide neatly along two dimensions: 1) belief-confirming consultations (BCC) that indicate an individual’s desire to seek guidance from sources that accord with their existing religious beliefs, and 2) belief-threatening consultations (BTC), which challenge one’s established religious worldview.[4] The respondents in our sample clearly preferred the BCC route when doubts arose. In particular, solitary engagement, either through prayer or turning to Islam’s primary sources, was the likeliest path our respondents took when faced with deep concerns about their faith.
Notably for our purposes, a clear majority reported being at least somewhat likely to talk to an imam, which gives us further confidence that the institutional leaders’ experiences in “Modern Pathways” align well (if perhaps not perfectly) with the broader American Muslim reality. Moreover, the tendency to turn to sources that reinforce faith, coupled with the overall lower levels of current doubt when compared to doubt at any time in the past, indicates that crises of faith are not irreversible and that interventions (whether personal or interpersonal) can effectively address certain apprehensions.
At the other end of the graph, although the tallies are significantly lower, there are nonetheless sizeable portions of the sample that explore doubts via a more contentious pathway. No fewer than a quarter of our respondents were at least somewhat likely to react to their doubt by seeking sources of guidance outside of Islam or otherwise resolving not to be bound to the goal of maintaining their faith. This is not to say, however, that a relinquishing of Islam is the endpoint of all who travel down this road. Indeed, some who make this journey may ultimately renew or even strengthen their belief. Such was the case, for example, with Yaqeen’s own Sheikh Omar Suleiman who, in a recent interview, related how he himself chose a more skeptical response to doubt in his youth and eventually ended up more certain in his faith as a result.[5]
The Correlates of Doubt
Methodology
In addition to descriptive data, we can further probe doubt in the American Muslim community through multivariate analysis. This statistical technique allows us to isolate the effect of key variables while holding other potentially influential factors constant. The reader should be careful, however, not to infer causality from the findings highlighted below. Since our data do not allow us to establish the presence of one variable prior to another, we cannot claim that our analyses reveal causal relationships. For example, if we find that lower religiosity (however defined) is associated with higher levels of doubt, our data cannot say whether the former is the cause or the effect of the latter, or whether some third factor influences both.
The outcome variable in each of the statistical models is the respondent’s total level of doubt over time. This construct is captured by adding each respondent’s answers to the questions in the main doubt scale, where “Not at all” is coded as a “0” and “A great deal” is coded as a “4.” For ease of interpretation, this sum is then rescaled from 0-1, where “0” and “1” represent the lowest and highest possible levels of total reported doubt, respectively. The x-axis in each of the remaining figures thus represents an increase or decrease on this scale.
The same rescaling was carried out for each of the explanatory variables in the model. This means that the results reported in the graphs correspond to the difference in the doubt scale when comparing the highest value of the independent variable to its lowest value. So, for example, if the effect of Age in a model is negative and statistically significant, that means that the oldest respondents in the sample, on average, report lower levels of doubt than the youngest respondents.
We tested three sets of potential correlates. First, a standard demographic model analyzed the impact of Age, Sex, Education, Convert status, and race/ethnicity (African-American, Asian, and Arab, all in reference to White respondents). Each of these items were subsequently included as controls in the remaining models. The second substantive set of variables gauged various aspects of religiosity: Mosque Attendance, Frequency of Prayer, Quran Literalism, and Importance of Religion in the Respondent’s Life.
A final suite of indices measured social and religious attitudes and experiences. Dogmatism gauges the extent to which one sees the world in nuanced versus black and white terms. Quest is an additive scale that taps the degree to which respondents perceive religion to be an unsettled search for truth as opposed to a fairly static set of beliefs. Religious Education tallies the respondents’ experience with various modes of religious learning as a youth. Religious Emphasis is an additive index that gauges the extent to which various aspects of religious life were stressed during the respondent’s youth. Finally, Religious Conservatism assesses the respondent’s attitudes toward the allowance or prohibition of several contentious issues in Islam, with a higher score indicating a tendency toward more prohibitive judgments. The full wording for each of these scales, as well as the religiosity items and demographic measures, is available in Appendix B.
Results
The results of the statistical analyses offer a mix of anticipated and surprising findings, both in terms of the items that registered as significant and those that did not. Figure 4 models the demographic correlates of doubt. Neither age nor education significantly impact religious doubt, which was a bit unexpected. More curious, however, is the differential effect of race: African-American, Asian, and Arab respondents all reported experiencing significantly lower levels of doubt than White respondents (the reference group). Any explanation of this finding would be speculative at this point, although the results are notably in line with extant research on religious conviction in the general public based on race/ethnicity.[6]
Figure 5 displays the effect of the various measures of religiosity. Of the variables in this set, only the importance of religion in the respondent’s life registers as statistically significant. Its effect is also quite substantial, as those for whom religion is very important are far lower in measured doubt than those who say religion is not at all important, with the difference being more than 25% of the total scale, on average. Of the remaining variables, the lack of a significant effect for Quran Literalism is quite surprising as this measure is typically associated with more entrenched beliefs.
The results in Figure 6 further evidence that conservative attitudes and experiences do not move the needle on doubt (at least, not in the expected direction). The result that stands out in terms of sheer magnitude is the impact of Quest. Fully 40% of the total doubt scale, on average, separates respondents who most view religion as a search for truth from those who least express this perspective. Perhaps the most remarkable finding across all the models, however, is that Dogmatism was positively associated with doubt. In effect, this means that those who are more close-minded tend to report a higher level of religious doubt. This result holds even when controlling for religiosity (see the fully specified model, Figure A, in the Appendix).
Similarly, growing up in an environment where religion was emphasized, rather than staving off doubt, seems to enhance it. In the full model (see Appendix C), this positive association crosses into statistical significance. This finding indicates that religious doubt may, in part, be a reaction to a strict upbringing (or at least one perceived to be so).
Continuing the trend, the effect of Religious Conservatism is similarly insignificant. Thus, those who have a hardline view on certain Muslim beliefs and practices are just as likely to express religious doubt as their more liberal counterparts. This result thus once again underscores that we should be cautious about simply assuming that certain types of individuals are more prone to have a crisis of faith.
Conclusion
What causes one person to doubt their faith and another to believe? Ultimately, from a theological perspective, the answer to this question rests in Divine decree. Through systematic analysis, however, we can better understand the observable attitudes and behaviors that tend to go hand in hand with religious doubt. The first stage of this project mapping American Muslim doubt highlighted three nodes around which doubts generally cluster: Moral and Social Concerns, Philosophical and Scientific Concerns, and Personal Trauma. These insights were subsequently used to inform a battery of questions in the 2017 Muslim American Attitudes Survey and provide large-n empirical data on the sources and correlates of doubt among American Muslims.
The preceding analysis featured three key sets of findings. First, the primary drivers of overall doubt appear to be the actions of Muslims, such as intolerance toward other faiths, rather than specific beliefs, such as the existence of God. Focusing in on the current sources of doubt among our sample, this trend largely maintains, but is likely a bit more informed by the negative socio-political events of recent years as “The bad things that people do in the name of religion” clearly tops the list. Second, the responses to doubt among our sample plainly evidence a tendency to turn to sources that would confirm one’s prior beliefs, such as the Quran or consultation with an imam, rather than threaten one’s faith, such as turning to sources that are expressly against Islam.
Third, the analysis highlighted a number of notable correlates of doubt, while also revealing some surprisingly inconsequential factors. Demographically, neither age, education, sex, nor convert status affected doubt, although Whites were curiously significantly higher on the doubt scale when compared to all other racial/ethnic groups. Remarkably, none of the typical measures of religiosity had a significant effect on reported levels of doubt, although the more respondents felt that religion was important in their life, the lower they tended to be on the doubt scale. By far the strongest correlate of doubt was the extent to which one regarded religion as an open-ended quest for truth, rather than a predefined set of beliefs. This result is not surprising when considering that the top reported sources of overall doubt underscored a keen aversion to dogmatism and hypocrisy.
These findings, alongside those from our qualitative study with community leaders, provide a fuller, empirical understanding of American Muslim doubt. As always, however, fruitful avenues remain for future research. One clear route would be to survey those individuals who have left Islam. It remains unclear whether the respondents in our sample, who self-identify as Muslim, are systematically different in their views on doubt than those who no longer count themselves as adherents of Islam. Understanding if and when the opinions of these two groups diverge would give us greater insight into what transforms doubt to disbelief.
A second useful extension of this project would be to incorporate experimental methods to establish a more causal mapping of American Muslim doubt. The application of this methodology will require an innovative design and will naturally bump up against logistical limitations. One possible way to bridge this causal gap would be to test whether increasing the salience of certain issues has a measurable effect on one’s certainty in Islam. Does, for example, informing respondents about an incident of spiritual abuse, intolerance, or hypocrisy affect their propensity to express doubt and, if so, is this effect constant across different demographics and levels of religiosity?
Much work remains to be done in this domain. In the meantime, the analysis presented here and in “Modern Pathways” offers community leaders, students, and educators rare, empirically-based insights into the issues affecting American Muslim faith.
Appendix A – Sampling Methodology
The online sample for this study was obtained through Qualtrics, a survey research firm. Respondents from numerous survey panels were invited to participate in the study if they had previously indicated that they were Muslim. The invitation itself was generic and made no mention of the specific topic. Incentives for participation are most often in the form of “points” that can be redeemed for gift cards, miles for airline loyalty programs, etc. This double opt-in process took place between August 25th and October 16th, yielding a total of 630 respondents. For both the descriptive and inferential analyses, the final sample is weighted to the Pew (2017) data on American Muslims’ age, education, and sex.
Appendix B – Key Variable Measures
Sex: What is your sex?
Age: Respondents were asked to input their age in years and were subsequently coded according to the following categories:
Education: What is the highest level of education you have completed or the highest degree you have received?
Race/Ethnicity: Which of the following categories best represents your racial/ethnic background (choose all that apply):
Convert: Thinking about your childhood, would you say that you were raised to be… (anything other than “Muslim” was coded as “Convert”)
Mosque Attendance: Aside from weddings or funerals, how often do you go to a mosque?
Frequency of Prayer: In general, how often do you pray salah or namaz (formal prayer)?
Quran Literalism: Which of the following statements comes closest to your personal beliefs about the Quran?
Importance of Religion in Respondent’s Life: How important is religion in your life?
Quest (5-point; Strongly Disagree – Strongly Agree)
Religious Emphasis (4-point; Not at all, A little, A moderate amount, A great deal)
Dogmatism (5-point; Strongly Disagree – Strongly Agree)
Religious Education (1-point for each checked item)
Religious Conservatism (4-point; Absolutely allowed, Allowed for the most part (with exceptions), Prohibited for the most part (with exceptions), Absolutely prohibited)
Appendix C – Additional Models
Notes
[1] Youssef Chouhoud, “Modern Pathways to Doubt in Islam” (Dallas, TX: Yaqeen Institute for Islamic Research, 2016). https://yaqeeninstitute.org/en/youssef-chouhoud/modern-pathways-to-doubt-in-islam/
[2] Bob Altemeyer and Bruce Hunsberger, Amazing Conversions: Why Some Turn to Faith and Others Abandon Religion (Amherst, N.Y.: Prometheus Books, 1997).
[3] The items were adapted from Altemeyer and Hunsberger.
[4] For elaboration on these two approaches to religious doubt, see Bruce Hunsberger, Michael Pratt, and S. Mark Pancer, “A Longitudinal Study of Religious Doubts in High School and beyond: Relationships, Stability, and Searching for Answers,” Journal for the Scientific Study of Religion 41, no. 2 (2002): 255–66.
[6] Pew’s breakdown of their 2015 Religious Landscape Study by race/ethnicity evidences a number of areas where White respondents are less likely to report adhering to fundamental religious practices or believe in core religious tenets than their Black, Latino, and Asian counterparts. See http://www.pewforum.org/religious-landscape-study/racial-and-ethnic-composition/
Authors